Volcanology - Fractional Crystallisation

Explain ‘fractional crystallisation’ and the roles it plays in the chemical evolution of magma and the explosivity of eruptions.

Fiona Fang, Trinity Hall, 11/2025, 1894 words

Fractional crystallisation is a fundamental process in magma differentiation, governing how a single parent magma can generate the wide spectrum of magma compositions observed on Earth (Oppenheimer, 2011). It refers to the progressive removal of early-formed crystals from a cooling melt, such that the residual liquid evolves chemically through the selective partitioning of elements between solid and liquid phases. This essay examines, in sequence, (1) the thermodynamic basis of crystallisation as outlined by Bowen’s reaction series; (2) how fractional crystallisation modifies melt chemistry and rheology; (3) its effects on volatile behaviour and bubble dynamics; and finally critically evaluates (4) how these combined processes govern the variability in eruption explosivity.

Crystallisation in cooling magmatic bodies usually occurs within an inward-growing solidification front (Marsh, 2015). The sequence of mineral crystallisation, as described by Bowen’s reaction series, is controlled primarily by the thermodynamic stability of each mineral phase. As magma cools, the system continuously seeks the lowest-energy, most stable state. At any given temperature and pressure, the mineral phase that most effectively reduces the system’s Gibbs free energy will begin to crystallise. High-temperature minerals, such as olivine and calcium-rich plagioclase, have simple structures with strong ionic bonds that can withstand high thermal energy, making them stable early in the cooling process. As temperature falls, the melt becomes more polymerised, and minerals with increasingly complex, silicon-rich frameworks, such as biotite, feldspar, and quartz, become energetically favoured and begin to crystallise (Bowen, 1928). Minerals crystallise along two main pathways: discontinuous and continuous series. In the discontinuous series, the minerals formed at each step have a distinct silicate structure, from isolated silicate tetrahedra (nesosilicates) in olivine, single chains (inosilicates) in pyroxene, double chains (inosilicates) in amphibole, to sheets (phyllosilicates) in biotite. In contrast, the continuous series describes the gradual chemical evolution of plagioclase feldspars from calcium-rich to sodium-rich compositions within the same crystal structure (Earle, 2015). Notably, in natural systems, fractional crystallisation requires not only chemical differentiation but also physical separation between crystals and melt. In crystal-poor magmas, where all crystallisation is confined within the solidification front, crystals are effectively trapped and cannot segregate from the melt, producing little differentiation. By contrast, crystal-rich slurries containing early-formed phenocrysts can achieve true fractionation through mechanisms such as crystal settling, compaction, and mush-flow migration (Marsh, 2015).

Figure 1 Bowen’s reaction chain (Source: Physical Geology, Chapter 3.3, from Earle, 2015) https://open.maricopa.edu/physicalgeology/chapter/3-3-crystallization-of-magma/

Fractional crystallisation progressively alters the chemical composition of a magma. As each mineral forms, its component elements partition between solid and liquid according to distribution coefficients (D_i=C_s/C_l), where C_sand C_lare the concentrations of an element in the solid and liquid respectively. The evolving melt composition can be described by the Rayleigh fractionation equation, C_l=C_0 F^((D_i-1)), where C_0is the initial concentration and Fthe remaining melt fraction. Since early-formed mafic minerals preferably incorporate Mg, Fe, and Ca, their removal depletes the melt in these cations while enriching it in Si, Al, Na, and K. The melt therefore evolves along a liquid line of descent toward more silicic compositions.

By altering the component of the melt, the fractional crystallization process can further impact the melt’s atomic structure and viscosity. Viscosity is the measure of the internal resistance to flow in a substance when shear stress is applied (Giordano, Cas and Wright, 2024), and is fundamentally governed by the degree of polymerisation in the silicate network. Silicate melts are built from SiO₄ tetrahedra linked by bridging oxygens (BOs) (Mysen et al., 1982), while and alkaline-earth ions such as Ca²⁺, Mg²⁺ and Fe²⁺ are modifiers of such polymerisation network. This because the high cation field strength of latter allows them to preferentially bond with oxygen atoms, introducing non-bridging oxygens (NBOs) that bonds to one Si only, with its charge balanced by the network-modifying cation. This cation-oxygen bond is weaker than the Si-O-Si bond, depolymerising the network structure and lowering its activation energy (Lesher and Spera, 2015). According to the Arrhenius relation, η=η_0 exp⁡[(E_a+PV_a)/RT], a lower activation energy corresponds to low viscosity, characteristic for mafic magmas (Shaw, 1972).

As the magma evolves down the Bowen’s reaction chain, the progressive removal of network-modifying cations lowers the abundance of NBOs and thereby increases the activation energy (E_a) required for viscous flow, from about 200 kJ mol⁻¹ in mafic and ultramafic melts to over 400 kJ mol⁻¹ in more silicic compositions (Lesher & Spera, 2015). This structural change explains why rhyolitic melts can be up to 10¹⁴ times more viscous than ultramafic counterparts such as komatiites or kimberlites (Dingwell 1998; Spera 2000; Giordano et al. 2008; Takeuchi, 2011, AGU). For instance, the rhyolitic melt (77.8 wt % SiO₂) from the 1934 Kikai lava dome exhibits a viscosity of η = 6.3 log Pa s (Ono, 1982; Saito et al., 2001, 2002), whereas the basaltic melt (48.8 wt % SiO₂) from the 1959 Kīlauea eruption is far more fluid, with η = 1.2 log Pa s (Macdonald & Katsura, 1961; Murata & Richter, 1966; Wallace & Anderson, 1998). Viscosity is the key control on volcanic eruption explosiveness, as it governs how efficiently gases can escape from magma. The underlying mechanism is examined in the following section on volatile and bubble dynamics.

Fractional crystallisation influences not only the melt composition but also the volatile content of magma. Again, according to the Rayleigh fractionation law, the concentration of a volatile component ibetween solid and liquid phases is governed by its partition coefficient (D_i=C_s/C_l). Because volatiles are strongly incompatible in most anhydrous minerals (D_i≪1), their concentration in the residual melt increases exponentially as crystallisation proceeds. For instance, during the 2018 Lower East Rift Zone (LERZ) eruption of Kīlauea Volcano, extensive fractional crystallization was invoked to explain the high volatile content in the magma: as melt MgO contents decreased from 6 to 0.5 wt%, H2O content increased by 4-5 times, reaching 2 wt% at eruption (Wieser et al., 2022).

The enrichment of volatiles has a direct yet composition-dependent effect on magma viscosity. Firstly, dissolved water break Si–O–Si bridges and depolymerise the silica melt: the H+ ion exchanges with any cation that is not in a tetrahedral co-ordination site (e.g. Na+), thereby hydrolysing one of the tetrahedral co-ordinating oxygens to OH−. Small concentrations of water may lower the viscosity of magma by several orders of magnitude (see figure 3) (Spera 2000; Burnham 1979). Notably, at a given water content, the viscosity of different magma types can vary by several orders of magnitude, reaffirming that melt composition is the dominant control on viscosity. The ability of dissolved water to lower viscosity is also greater in silica-rich magmas than in silica-poor ones, simply because silicic melts contain more Si–O–Si linkages available for hydrolysis (Giordano, Cas & Wright, 2024). In contrast, dissolved CO2 tends to enhance polymerisation and therefore viscosity, although the exact mechanism is still under debate. While some argues it does so by forming CO3− complexes (Eggler and Rosenhauer 1978; Mysen et al. 1982), others attribute this to the formation of stabilised carbonate ions with free potassium (Bourgue and Richet 2001). The exact effects of other volatiles on magma viscosity are less well understood (Giordano, Cas and Wright, 2024).

Figure 2 Effect of dissolved water on magma viscosity for compositions ranging from komatiite to rhyolite (Source: Lesher and Spera 2015)

As volatiles dissolve into the melt, their concentration gradually approaches the saturation limit. Volatile solubility follow the empirical law C_l^sat=kP^n, where Cℓsat is the equilibrium concentration at saturation, P is pressure, and k and n are constants determined by volatile species and melt composition: As seen in Figure 3 (a) (b), for a given volatile and melt composition, CO₂ exhibits near-Henrian behaviour (n=1), whereas H₂O shows a strongly non-linear solubility that increases steeply with pressure (n<1) (Parfitt and Wilson, 2025). The solubility of water in rhyolite is considerably greater than that in basaltic magma (Figure 3a), while the solubility of CO2 in basaltic and rhyolitic magma is very similar, and is considerably less than the solubility of H2O in such magmas. The solubility of water is also controlled by the presence of other volatiles, especially CO2 (Williams and McBirney 1979; Wallace et al. 2015) and the presence of exchangeable cations which are not in tetrahedral co-ordination sites (Burnham 1979; Mysen et al. 1982).

Figure 3 Solubility conditions of H2O and CO2 in rhyolite and basalt magmas (source: Wallace et al. 2015)

When C_l>C_l^sat, the melt becomes supersaturated, and a gas phase nucleates—a process termed secondary boiling. Interestingly, fractional crystallization can help nucleation process because gas can use the irregular surface of crystals as nucleation sites to minimize the effect of surface tension, known as heterogeneous nucleation (Parfitt and Wilson, 2008). Once formed, the bubble can grow by diffusion, decompression and coalescence. Diffusion dominates at small bubble sizes, where gas molecules migrate along concentration gradients established around growing bubbles. During magma ascent, decompression drives further expansion according to Boyle’s law (PV=”constant” ), as reduced pressure (P) increases bubble volume (V). Coalescence becomes important in slowly ascending magma, where differential rise velocities cause smaller bubbles to be entrained in the wakes of larger ones, leading to collisions and merging. This process can generate transient overpressure and cause discrete explosive bursts, typical of Strombolian and Vulcanian activity (Parfitt & Wilson, 2008).

In scenarios where volatiles reach saturation and bubbles form, the key control on explosivity is the efficiency of gas escape. Fractional crystallisation enriches the melt in silica and thereby increases viscosity, while the removal of dissolved water during second boiling further enhances this effect. High viscosity impedes bubble ascent and coalescence, allowing pressure to accumulate within the magma chamber until the internal overpressure exceeds the magma’s tensile strength, triggering brittle fragmentation and explosive discharge (Zhang, 1999). The 2015 Calbuco eruption illustrates this process, where crystallisation-induced second boiling and a sealed system led to over-pressurisation and sub-Plinian explosivity (Arzilli et al., 2019). However, rapid vesiculation can alternatively generate efficient bubble coalescence, enabling a transition from closed- to open-system degassing that permits permeable gas escape, as is commonly observed in basaltic magmas (Namiki & Manga, 2008). The presence of crystals may further facilitate gas escape by lowering the percolation threshold for bubble connectivity, enhancing conduit-scale permeability even at low gas fractions (Colombier et al., 2021; Collombet et al., 2021). Under such conditions, eruptions are more likely effusive than explosive. Finally, eruptive style can also vary spatially and temporally within the same volcano. As documented by Wieser et al (2022), during the 2018 Lower East Rift Zone eruption of Kīlauea, magma from Fissure 17 had undergone extensive fractional crystallisation (~0.5 wt% MgO, ~2 wt% H₂O) and produced Strombolian-to-explosive activity, whereas less evolved magma from adjacent fissures erupted effusively. This contrast exemplifies how fractional crystallisation process controls eruptive behaviour by impacting the degree of melt evolution and the openness of the degassing system.

To sum up, fractional crystallisation links the chemical differentiation of magma to its eruptive behaviour. By removing early-formed crystals, it enriches the melt in silica and volatiles and raises viscosity, creating conditions that heighten the potential for overpressure and explosive discharge. Yet explosivity is not inevitable. Where crystallisation and vesiculation create permeable networks, gases escape efficiently and eruptions remain effusive, as at Kīlauea; when the system is sealed, volatile retention drives fragmentation, as at Calbuco. Fractional crystallisation thus enables explosivity but does not determine it. Its role is to modulate the balance between volatile supply and degassing efficiency, making chemical evolution and eruption dynamics two expressions of the same process.

##Bibliography

Arzilli, F., La Spina, G., Burton, M.R., Polacci, M., Le Gall, N., Hartley, M.E., Di Genova, D., Cai, B., Vo, N.T., Bamber, E.C. and Nonni, S., 2019. Magma fragmentation in highly explosive basaltic eruptions induced by rapid crystallization. Nature Geoscience, 12(12), pp.1023-1028.

Arzilli, F., Morgavi, D., Petrelli, M., Polacci, M., Burton, M., Di Genova, D., Spina, L., La Spina, G., Hartley, M.E., Romero, J.E. and Fellowes, J., 2019. The unexpected explosive sub-Plinian eruption of Calbuco volcano (22–23 April 2015; southern Chile): Triggering mechanism implications. Journal of Volcanology and Geothermal Research, 378, pp.35-50.

Bertagnini, A., Métrich, N., Landi, P. and Rosi, M., 2003. Stromboli volcano (Aeolian Archipelago, Italy): An open window on the deep‐feeding system of a steady state basaltic volcano. Journal of Geophysical Research: Solid Earth, 108(B7).

Bowen, N.L., 1928. The evolution of the igneous rocks. Princeton University Press.

Bourgue, E. and Richet, P., 2001. The effects of dissolved CO2 on the density and viscosity of silicate melts: a preliminary study. Earth and Planetary Science Letters, 193(1-2), pp.57-68.

Burnham, C.W., 1979. Importance of volatile constituents. Evolution of igneous rocks.

Colombier, M., Vasseur, J., Houghton, B.F., Cáceres, F., Scheu, B., Kueppers, U., Thivet, S., Gurioli, L., Montanaro, C., Soldati, A. and Di Muro, A., 2021. Degassing and gas percolation in basaltic magmas. Earth and Planetary Science Letters, 573, p.117134.

Collombet, M., Burgisser, A., Colombier, M. and Gaunt, E., 2021. Evidence for deep gas loss in open volcanic systems. Bulletin of Volcanology, 83(2), p.7.

Dingwell, D.B., 1998. Recent experimental progress in the physical description of silicic magma relevant to explosive volcanism. Geological Society, London, Special Publications, 145(1), pp.9-26.

Earle, S., 2015. Physical geology. BCcampus.

Eggler, D.H. and Rosenhauer, M., 1978. Carbon dioxide in silicate melts; II, Solubilities of CO 2 and H 2 O in CaMgSi 2 O 6 (diopside) liquids and vapors at pressures to 40 kb. American Journal of Science, 278(1), pp.64-94.

Giordano, D., Russell, J.K. and Dingwell, D.B., 2008. Viscosity of magmatic liquids: a model. Earth and Planetary Science Letters, 271(1-4), pp.123-134.

Giordano, G., Cas, R. and Wright, J.V., 2024. Properties of Magmas: Relevance to Eruption Processes and Lava Flow. In Volcanology: Processes, Deposits, Geology and Resources (pp. 37-74). Cham: Springer International Publishing.

Lesher, C.E. and Spera, F.J., 2015. Thermodynamic and transport properties of silicate melts and magma. In The encyclopedia of volcanoes (pp. 113-141). Academic Press.

MacDonald, G.A. and Katsura, T., 1961. Variations in the lava of the 1959 eruption in Kilauea Iki.

Marsh, B.D., 2015. Magma chambers. In The encyclopedia of volcanoes (pp. 185-201). Academic Press.

Murata, K.J. and Richter, D.H., 1966. Chemistry of the lavas of the 1959-60 eruption of Kilauea Volcano, Hawaii (No. 537-A).

Mysen, B.O., Virgo, D. and Seifert, F.A., 1982. The structure of silicate melts: implications for chemical and physical properties of natural magma. Reviews of Geophysics, 20(3), pp.353-383.

Namiki, A. and Manga, M., 2008. Transition between fragmentation and permeable outgassing of low viscosity magmas. Journal of Volcanology and Geothermal Research, 169(1-2), pp.48-60.

Ono, K., T. Soya, and T. Hosono (1982), Geology of the Satsuma-Io-Jima district [in Japanese with English abstract], Quadrangle Ser., scale 1:50000, 80 pp., Geol. Surv. Jpn., Tokyo.

Oppenheimer, C., 2011. Eruptions that shook the world. Cambridge University Press.

Parfitt, E., Wilson, L. and Kerber, L., 2025. Fundamentals of physical volcanology. John Wiley & Sons.

Preston, R.M.F., 1982. H. Williams, and AR McBirney. Volcanology. San Francisco (Freeman, Cooper & Co.), 1979. 398 pp., 233 figs. Price $33.50. Mineralogical Magazine, 46(340), pp.405-406.

Richter, D.H. and Murata, K.J., 1966. Petrography of the lavas of the 1959-60 eruption of Kilauea Volcano, Hawaii (No. 537-D).

Saito, G., Kazahaya, K., Shinohara, H., Stimac, J. and Kawanabe, Y., 2001. Variation of volatile concentration in a magma system of Satsuma-Iwojima volcano deduced from melt inclusion analyses. Journal of Volcanology and Geothermal Research, 108(1-4), pp.11-31.

Saito, G., Stimac, J.A., Kawanabe, Y. and Goff, F., 2002. Mafic-felsic magma interaction at Satsuma-Iwojima volcano, Japan: Evidence from mafic inclusions in rhyolites. Earth, planets and space, 54(3), pp.303-325.

Shaw, H.R., 1972. Viscosities of magmatic silicate liquids; an empirical method of prediction. American journal of Science, 272(9), pp.870-893.

Spera, F.J., 2000. Physical properties of magmas. Encyclopedia of volcanoes.

Takeuchi, S., 2011. Preeruptive magma viscosity: An important measure of magma eruptibility. Journal of Geophysical Research: Solid Earth, 116(B10).

Wallace, P.J. and Anderson Jr, A.T., 1998. Effects of eruption and lava drainback on the H2O contents of basaltic magmas at Kilauea Volcano. Bulletin of Volcanology, 59(5), pp.327-344.

Wallace, P.J., Plank, T., Edmonds, M. and Hauri, E.H., 2015. Volatiles in magmas. In The encyclopedia of volcanoes (pp. 163-183). Academic Press.

Wieser, P.E., Edmonds, M., Gansecki, C., Maclennan, J., Jenner, F.E., Kunz, B., Antoshechkina, P., Trusdell, F., Lee, R.L. and EIMF, 2022. Explosive Activity on Kīlauea’s Lower East Rift Zone Fueled by a Volatile‐Rich, Dacitic Melt. Geochemistry, Geophysics, Geosystems, 23(2), p.e2021GC010046.

Zhang, Y., 1999. A criterion for the fragmentation of bubbly magma based on brittle failure theory. Nature, 402(6762), pp.648-650.